The invention relates generally to light microscopy and, more particularly, to automated techniques of analyzing cytochemical and immunohistochemical staining.
In current cellular imaging systems, the area of a stained slide containing a stained cellular specimen is scanned by automated microscopy systems. The entire cellular specimen area of the slide is imaged with a series of field of view images. For further analysis, each field of view image must be separately analyzed to locate complete candidate objects of interest within the field of view image. This approach may be acceptable for cellular objects or clusters that are not so large that they extend beyond the field of view width or length of an image. Often, for both automated and manual analysis, only a single field of view is analyzed for morphological characteristics, so that the context of the analysis is limited to that field of view on a single slide.
A method for automated image analysis of a biological specimen is disclosed. A biological specimen is stained and counterstained for a specific marker or in the instance of immunohistochemistry or in situ hybridization, the marker is detectably labeled. Such labels include enzyme, radioisotopes, fluorescence or other labels well known in the art. The sample is then scanned at a plurality of positions by a photoimaging system to acquire an image. Adjacent positions are then used to reconstruct and provide a full picture image of the whole sample. A reconstructed whole sample image may then be further processed to identify coordinates that may have an object or area of interest. These coordinates are automatically selected for a candidate object of interest. In a preferred embodiment, a low magnification image of the candidate objects of interest is automatically obtained. Preferably the image is a color digital image. The candidate object of interest pixels in each sample are automatically morphologically processed to identify artifact pixels and the remaining candidate object of interest pixels in the sample not identified as artifact pixels. The apparatus obtaining the low magnification image is adjusted to a higher magnification, to acquire a higher magnification image of the sample, at the location coordinates corresponding to the low magnification image, for each candidate object or area of interest. Pixels of the higher magnification image in the first color space are automatically transformed to a second color space to differentiate higher magnification candidate objects of interest pixels from background pixels and identify, at higher magnification, objects of interest from the candidate object of interest pixels in the second color space. Thus, the pathologist or technician can identify whether the candidate object of interest has been specifically stained for the marker of interest, or counterstained, or both specifically stained and counterstained.
An automated cellular image method for analyzing a biological specimen, that has consecutively been stained with hematoxylin and eosin (H/E) on one tissue section and by one or several immunohistochemistry (IHC) and/or in situ hybridization (ISH) methods on parallel tissue sections, is also disclosed as a particular embodiment. To identify a structure in tissue that cannot be captured in a single field of view image or a single staining technique, the invention provides a method for histological reconstruction to analyze many fields of view on potentially many slides simultaneously. The method couples composite images in an automated manner for processing and analysis. A slide on which is mounted a biological specimen stained to identify structure of interest is supported on a motorized stage. An image of the biological specimen is generated, digitized, and stored. As the viewing field of the objective lens is smaller than the entire biological specimen, a histological reconstruction is made. These stored images of the entire tissue section may then be placed together in an order such that the H/E stained slide is paired with the immunohistochemistry slide so that analysis of the images may be performed simultaneously by the pathologist. The images may be superimposed or viewed as adjacent images.
In one embodiment, the invention provides a method for automated image analysis of a biological specimen by providing a biological sample to be analyzed, automatically scanning the sample at a plurality of coordinates, automatically obtaining an image at each of the coordinates, reconstructing an image of the sample from each individual image to create a reconstructed image and processing the reconstructed image to identify a candidate object or area of interest. In another embodiment, the invention provides a method for histological reconstruction. In this method a sample of a biological specimen is divided into a number of subsamples. Each subsample is processed with a stain, counterstain, immunohistochemical technique, in situ hybridization technique, or a combination thereof. Each sample is then scanned and an image is obtained from each of the samples. The images are then reconstructed such that a first image is paired with a consecutive-corresponding sample image for identification of objects or areas of interest.
In yet another embodiment, the invention provides a computer program residing on a computer-readable medium, for automated image analysis of a biological specimen. The computer program comprises instructions for causing a computer to process a sample by scanning the sample at a plurality of coordinates, obtaining an image at each of the coordinates, reconstructing the sample from the individual images to create a reconstructed sample, identifying a coordinate of a candidate object or area of interest in the reconstructed image and acquiring an image at the coordinate obtained from the reconstructed image.